Primary brain tumors, including malignant gliomas and ependymomas, are second only to leukemia as a cause of childhood cancer. Advances in chemotherapy and combined modality regimens, though dramatically successful in the treatment of other pediatric malignancies, have not been translated into effective therapy for most pediatric brain tumors. Although the role of chemotherapy in the treatment of malignant gliomas and ependymomas remains poorly defined, recent clinical and laboratory studies support the activity of the classical bifunctional alkylating agents against these tumors. Nevertheless, substantial increases in patient survival as a result of adjuvant use of these agents remain to be demonstrated, and an understanding of the mechanisms responsible for drug failure is critical for the design of optimal chemotherapeutic intervention. Successful establishment of cell line and xenograft models of childhood high grade glioma and ependymoma now provide the biological tools to facilitate an understanding of alkylator resistance in these tumors. Resistance to alkylating agents, including cyclophosphamide and melphalan, is multifactorial, with a diverse spectrum of mechanisms observed in murine and human neoplasia. Mechanisms of resistance to cyclophosphamide include increased aldehyde dehydrogenase activity., increased glutathione-S- transferase activity, elevated levels of glutathione, and a presently undefined mechanism in medulloblastoma. Similarly, mechanisms of resistance to malphalan include decreased cellular transport, increased intracellular glutathione levels protective of critical cellular targets, cellular detoxification and enhanced capacity to repair damaged DNA. These studies may not be relevant to the mechanisms of resistance operational in childhood malignant glioma and ependymoma. The hypothesis of this proposal is: definition and modulation/bypass of alkylator resistance in childhood high grade glioma and ependymoma will allow selection of alkylator regimens active in the treatment of these tumors and increase survival of children with these neoplasms. The specific aims of this proposal are: 1) To continue to establish childhood high grade glioma and ependymoma cell lines and transplantable xenografts in athymic mice with de novo clinical, acquired clinical and laboratory-generated cyclophosphamide and melphalan resistance. 2) To define the mechanisms of resistance to cyclophosphamide and melphalan of childhood high grade glioma and ependymoma cell lines and xenografts with de novo, acquired clinical, and laboratory-generated resistance. 3) To define modulation effective in bypassing/reversing cyclophosphamide and melphalan resistance in childhood high grade glioma and ependymoma cell lines and xenografts and 4) To define the role of L- amino acid oxidase-mediated depletion of plasma large neutral amino acids to enhance delivery and activity of melphalan in the treatment of subcutaneous and intracranial childhood high grade glioma and ependymoma xenografts in athymic mice.